Faulting and earthquake triggering during the 1783 Calabria seismic sequence E. Jacques, 1, * C. Monaco, 2 P. Tapponnier, 1 L. Tortorici 2 and T. Winter 3 1 Laboratoire de Tectonique et Me ´canique de la Lithosphe `re UMR CNRS 7578, Institut de Physique du Globe de Paris, France. E-mails: [email protected]; [email protected]2 Istituto di Geologia e Geofisica, Universita ` di Catania, Italy. E-mails: [email protected]; [email protected]3 Bureau de Recherche Ge ´ologique et Minie `re, Orle ´ans, France Accepted 2001 May 17. Received 2001 April 19; in original form 1999 March 26 SUMMARY Between the 1783 February 5 and 1783 March 28, five earthquakes struck the southern part of Calabria. The main shock (February 5) and the first aftershock (February 6) devastated the region ENE of the Messina Strait. The greatest damage occurred along the foot of the Aspromonte Mountains south of San Giorgio Morgeto, and along the Tyrrhenian coast south of Palmi. A surface break about 18 km long, with several feet of downthrow to the west, formed along the Cittanova (Santa Cristina) Fault as a result of the main shock. On February 7, a third large shock ruined villages at the foot of the Serre Mountains north of San Giorgio Morgeto. Morphological and structural evidence, combined with a reassessment of observations made at the time of the earthquakes, suggest that these three shocks were shallow (j20 km) and related to slip on the west- dipping, NE-striking Cittanova–Sant’Eufemia, Palmi–Scilla and Serre normal faults, respectively, which juxtapose the basement of the Aspromonte and Serre mountains with the Pleistocene deposits of the Gioia Tauro and Mesima basins, and border the Palmi coastal high. The three faults belong to an active rift that stretches from northern Calabria to offshore the Ionian coast of Sicily. The spatial coupling between the 1783 events is investigated by resolving changes of Coulomb failure stress. The main shock (1783 February 5, My7), on the Cittanova and Sant’Eufemia faults, increased that stress by several bars on the Scilla Fault, triggering the 1783 February 6 earthquake (My6.5). The cumulative effect of these two shocks was to raise the Coulomb stress by more than 1 bar on the SW part of the Serre Fault, which was subsequently the site of the 1783 February 7 shock (My6.5). In turn, the first three events increased the stress by about 1 bar on the NE part of this latter fault, leading to the 1783 March 1 shock (My5.7). The gap between the 1783 February 7 and 1783 March 1 events may be related to the previous occurrence of an earthquake 124 yr before (1659 November 5, My6), which had already released stress locally. The occurrence of the last 1783 event (28 March) is not as simply accounted for by Coulomb modelling, in part because it remains unclear which fault slipped and how deep this event was. Overall, the 1783 sequence increased the Coulomb failure stress by several bars south of the Messina Strait and north of the epicentral region of the 1693 SE Sicily (Catania–Noto) earthquakes. 125 yr later, this same region was the site of the 1908 Messina earthquake, also a normal faulting event. Our study thus provides one convincing example in which Coulomb stress modelling brings insight into the spatial dynamics of seismic sequences. Key words: crustal deformation, earthquakes, Italy, normal faulting, stress distribution. * Now at: Equipe de Tectonique active et Pale ´osismologie, UMR CNRS 7516, Institut de Physique du Globe de Strasbourg, France. E-mail: [email protected]Geophys. J. Int. (2001) 147, 499–516 # 2001 RAS 499
18
Embed
Faulting and earthquake triggering during the 1783 ...vigny/articles/calabre_1783.pdf · Faulting and earthquake triggering during the 1783 Calabria ... the foot of the Aspromonte
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Faulting and earthquake triggering during the 1783 Calabriaseismic sequence
E. Jacques,1,* C. Monaco,2 P. Tapponnier,1 L. Tortorici2 and T. Winter31 Laboratoire de Tectonique et Mecanique de la Lithosphere UMR CNRS 7578, Institut de Physique du Globe de Paris, France.
Accepted 2001 May 17. Received 2001 April 19; in original form 1999 March 26
SUMMARY
Between the 1783 February 5 and 1783 March 28, five earthquakes struck the southernpart of Calabria. The main shock (February 5) and the first aftershock (February 6)devastated the region ENE of the Messina Strait. The greatest damage occurred alongthe foot of the Aspromonte Mountains south of San Giorgio Morgeto, and along theTyrrhenian coast south of Palmi. A surface break about 18 km long, with several feet ofdownthrow to the west, formed along the Cittanova (Santa Cristina) Fault as a result ofthe main shock. On February 7, a third large shock ruined villages at the foot of the SerreMountains north of San Giorgio Morgeto. Morphological and structural evidence,combined with a reassessment of observations made at the time of the earthquakes,suggest that these three shocks were shallow (j20 km) and related to slip on the west-dipping, NE-striking Cittanova–Sant’Eufemia, Palmi–Scilla and Serre normal faults,respectively, which juxtapose the basement of the Aspromonte and Serre mountainswith the Pleistocene deposits of the Gioia Tauro and Mesima basins, and border thePalmi coastal high. The three faults belong to an active rift that stretches from northernCalabria to offshore the Ionian coast of Sicily. The spatial coupling between the 1783events is investigated by resolving changes of Coulomb failure stress. The main shock(1783 February 5, My7), on the Cittanova and Sant’Eufemia faults, increased thatstress by several bars on the Scilla Fault, triggering the 1783 February 6 earthquake(My6.5). The cumulative effect of these two shocks was to raise the Coulomb stress bymore than 1 bar on the SW part of the Serre Fault, which was subsequently the site ofthe 1783 February 7 shock (My6.5). In turn, the first three events increased the stressby about 1 bar on the NE part of this latter fault, leading to the 1783 March 1 shock(My5.7). The gap between the 1783 February 7 and 1783 March 1 events may be relatedto the previous occurrence of an earthquake 124 yr before (1659 November 5, My6),which had already released stress locally. The occurrence of the last 1783 event(28 March) is not as simply accounted for by Coulomb modelling, in part becauseit remains unclear which fault slipped and how deep this event was. Overall, the 1783sequence increased the Coulomb failure stress by several bars south of theMessina Straitand north of the epicentral region of the 1693 SE Sicily (Catania–Noto) earthquakes.125 yr later, this same region was the site of the 1908 Messina earthquake, also a normalfaulting event. Our study thus provides one convincing example in which Coulombstress modelling brings insight into the spatial dynamics of seismic sequences.
Key words: crustal deformation, earthquakes, Italy, normal faulting, stress distribution.
*Now at: Equipe de Tectonique active et Paleosismologie, UMR CNRS 7516, Institut de Physique du Globe de Strasbourg, France.
Figure 1. Seismotectonic map of the Calabrian Arc and eastern Sicily. Faults are mainly normal (barbs on downthrown block) and Pleistocene–
Holocene in age. Crustal seismicity (H<35 km) since 1000 AD (Postpischl 1985; Boschi et al. 1995) and focal mechanisms of instrumental
earthquakes (Gasparini et al. 1982; Cello et al. 1982; Anderson & Jackson 1987; Amato et al. 1995) are shown. Numbers refer to fault segments:
1, Cittanova; 2, Sant’Eufemia; 3, Scilla; 4, Galatro; 5, Serre. The events of the 1783 earthquake sequence are represented as black dots. Boxes indicate
the locations of Figs 8 and 9. Inset shows kinematic slip indicators on fault segments 1 and 2.
500 E. Jacques et al.
# 2001 RAS, GJI 147, 499–516
rift zone’ (Monaco et al. 1997). Several of the faults cut and
offset Upper Pleistocene to Holocene sediments along escarp-
ments with young morphology, which bound steep range fronts
(Catena costiera, Aspromonte, Serre, Peloritani). The high level
of crustal seismicity is attested by numerous instrumental and
historical earthquakes since 1000 AD. (Postpischl 1985; Boschi
et al. 1995) with magnitudes of up to 7.1 and intensities of up
to XI–XII MCS. Tectonic studies (e.g. Tapponnier et al. 1987;
Mazzuoli et al. 1995; Tortorici et al. 1995; Monaco et al. 1995,
1997) and fault plane solutions (Gasparini et al. 1982; Cello
et al. 1982; Anderson & Jackson 1987) show that, overall, the
direction of extension along the Quaternary Siculo-Calabrian
rift zone is roughly ESE–WNW.
In 1783, the southern part of Calabria was struck by one of
the most devastating earthquake sequences in Western Europe.
Five large shocks occurred along a zone about 100 km long in
less than two months. In this paper, we first identify the faults
that ruptured during the earthquakes. For this, we combine
a reassessment of the available historical information with
geological and geomorphological field evidence supported by
the analysis of 1:33 000 scale aerial photographs and SPOT
satellite images. After elucidating the pattern of interacting
fault segments that broke, we try to understand the spatial
coupling between the events of the sequence by resolving the
changes of Coulomb failure stress caused by each earthquake
rupture using a simple dislocation model in an elastic half-space
(King et al. 1994). In doing this, we follow studies performed in
California (e.g. Reasenberg & Simpson 1992; Stein et al. 1992;
King et al. 1994), in Afar (Jacques 1995; Jacques et al. 1996)
and in the Apennines (Nostro et al. 1997) that have shown that
modelling Coulomb failure stress changes induced by one or
more events can clarify the distribution of future events. The
stress changes are calculated on faults with given strikes, or on
optimally oriented faults, given the regional stress field. We
finally broaden the regional scope of our study by including
in the modelling all the large earthquakes that have occurred in
eastern Sicily and southern Calabria over the last 350 yr.
GEOLOGICAL AND STRUCTURALFRAMEWORK OF SOUTHWESTERNCALABRIA
The Aspromonte and Serre mountains (maximum elevation
1956 m at Montalto) form the structural backbone of SW
Calabria. They are composed of uplifted metamorphic base-
ment (chiefly gneisses and granites). The basement is a pile
of several thrust sheets, involving Cretaceous sediments in
the eastern, external part (e.g. Ogniben 1973). The mountains
slope rather gently towards the Ionian Sea to the east, but their
western edge is steeper and controlled by normal faults. The
faults dip west and bound a major Plio-Pleistocene trough
(Gioia Tauro and Mesima basins) (Fig. 2). The southern part
of the Gioia Basin is separated from the Tyrrhenian Sea by the
Palmi high (maximum elevation 593 m), along which basement
identical to that in the Aspromonte has been uplifted and tilted
by movement on another west-dipping, offshore normal fault
(Figs 2a and b). On this high, the basement is unconformably
covered by patches of calcareous sandstones and white marls of
Tortonian to Early Pliocene age (Figs 2a and b). The Gioia
Basin is filled by about 600 m of Plio-Pleistocene marine sedi-
ments (Fig. 2b). The basal deposits, visible near Sant’Eufemia
and San Giorgio Morgeto, are cross-bedded sands and
calcarenites with conglomeratic lenses of Late Pliocene–Early
Pleistocene age (#2 Ma), about 70 m thick (beach deposits,
Fig. 2a). They are overlain by a clay and silt sequence several
hundred metres thick. Pumice-rich horizons are found in the
upper part of the sequence. Bathyal and circalitoral micro-
faunal assemblages imply that, as for the analogous sequence
found south of Villa San Giovanni (Barrier 1987), the clays
were deposited between the Early and Middle Pleistocene
(1.8–0.5 Ma) in water depths of around 400 m. In much of the
Gioia Basin, 70 m cross-bedded beach sands cap, with slight
angular unconformity, the marine sequence. These sands onlap
the basement in the Delianova area (Fig. 2a). Near Molocchio
and Oppido Mamertina, coarser conglomeratic lenses inter-
finger with the sands, implying synsedimentary tectonic activity
along the edge of the Aspromonte (Fig. 2b). The upper beach
sands extend laterally to the Tyrrhenian coastal deposits, in
which the typical Strombus Bubonius fauna is found (Bonfiglio
& Berdar 1986). Given the probable westward progradation of
these deposits and westward migration of the coastline, how-
ever, the age of the upper beach sands in the Gioia Basin may
only be estimated to be between the upper Middle Pleistocene
(200 ka) and the Late Pleistocene (100 ka). Middle to Late
Pleistocene deposits are also found south of Villa San Giovanni
(Dumas et al. 1979; Barrier & Keraudren 1983; Barrier 1987)
(Fig. 2a), but with a different facies (Gilbert-type fan delta
conglomerates). Between Villa San Giovanni and Gioia Tauro,
as many as nine to 10 levels of Pleistocene marine terraces step
up to about 600 m on the Palmi–Bagnara coastal high (Gignoux
25 /1143 i) (Vivenzio 1783), situated only 5 km southeast of
the crest of the Aspromonte and almost opposite Cittanova,
Molochio and Santa Cristina (Fig. 4), imply not only that
ground shaking intensities decreased dramatically across the
Aspromonte, but also that the strike of the southeastern limit
of the maximum intensity zone could not have been very
different from that shown. The position chosen for this limit in
Fig. 4 minimizes the surface of this zone. The relatively large
number of fatalities in Messina, which is located outside the
maximum intensity zone of Fig. 4, is probably mostly a con-
sequence of its large population, in addition to the architectural
weakening noted by Dolomieu. On the other hand, the low
fatality rates north of Galatro (Fig. 4), in spite of the occur-
rence of the shocks on February 7 and March 1, may be related
to the fact (usual in such circumstances and here mentioned by
several chroniclers) that most people, frightened by the wide-
ranging effects of the first two earthquakes, were then camping
out of their houses. Such an inference may not hold for the
shock on March 28, six weeks after the main February 5 event
and nearly 100 km northeast of the Messina strait, hence the
relatively large number of fatalities in Borgia (Table 1), which
are attributable to that shock only.
Table 1. List of towns and villages where 1783 earthquakes killed
more than 300 people, in order of decreasing number of casualties,
from Baratta’s (1901) table (pp. 284–288).
Locality Casualties Locality Casualties
Bagnara 3331 Radicena 756
Polistena 2261 Messina 617
Casalnuovo (Cittanova) 2017 Molochio 600
Terranova 1458 Varapodio 497
Scilla 1450 Sinopoli 382
Seminara 1370 Galatro 341
Cinquefrondi 1343 Borgia 332
San Giorgio Morgeto 1312 Paracorio 325
Oppido 1198 San Procopio 316
Palmi 0999 Iatrinoli 312
Sant’Eufemia di Sinopoli 0945 Tresilico 310
Santa Cristina d’Aspromonte 0760
504 E. Jacques et al.
# 2001 RAS, GJI 147, 499–516
Locations, intensities and inferred magnitudes of thevarious shocks
The observations compiled in Fig. 4 yield a clear macroseismic
picture of the first two shocks of the 1783 sequence. They
complement and give firmer basis to inferences made in pre-
vious reports (e.g. Baratta 1901) and by those who visited the
area at the time. Two features of the zone of maximum ground
shaking intensity in Fig. 4 are robust. First, the zone has two
main lobes connected by a narrow isthmus near Sant’Eufemia.
The smaller lobe roughly coincides with the Palmi high along
the Tyrrhenian coast. The larger lobe is a 5–10 km wide, 35 km
long strip of the Gioia Basin along the steep northwestern edge
of the Aspromonte. Second, shaking intensities seem to have
decreased rapidly from peak values causing complete collapse
of all constructions and fatality rates of 50 per cent or more
along the axial zones of these two lobes to much smaller
values, sparing more than 85 per cent of the population less
than 10 km away from such paroxysmal zones. These two
features imply that the two shocks had distinct hypocentres at
shallow crustal depths (<20 km). The most likely locations
of the two hypocentres are clearly under the eastern part of
the Gioia Basin (probably between Oppido and Molochio)
for the main February 5 shock, and off the Tyrrhenian coast
(probably between Scilla and Bagnara) for the February 6 shock.
Such hypocentres correspond to the sites inferred to have
38°20' N
38°40' N
38°00' N3°30' E 4°00' E
45 < %D
25 < %D < 45
15 < %D < 25
%D < 15
%D: fatality rate
N
10 km
surface rupture of 5/02/1783 event(Dolomieu, 1784)
IX
X
XI
X
X XI
XI
X-XI
Dinami
Acquaro
Mileto
Gerocarne
Soriano CalabroSorianello
Pizzoni
S. Nicola Da Crissa
S. Giorgio Morgeto
Cinquefrondi
Galatro
Canolo
Antonimina
Platì
Scilla
Bagnara
Seminara
Palmi
Terranova
Molochio
(Iatrinoli)
Taurianova(Radicena) Cittanova
PolistenaGioia Tauro
S. EufemiaS. Procopio
Sinopoli vecchio
S. Cristina d'Aspromonte
Oppido Mamertino
TresilicoCastellace
Varapodio
Messina
Reggiodi Calabria
Villa S. Giovanni
Delianuova
Arena
Capo Peloro
Capo Vaticano
M
A
R
E
T
I
R
R
E
N
O
M
A
R
E
I
O
N
I
O
B A
C I
N
O
D
I
G
I
O
I A
STRETTO DI M
ESSINA
A
S
P
R
O
M
T
ES
E
R
R
E
BA
CI N
O
DE
L
ME
SI
MA
1423m
1260m
1099m
989m 1276m
1143m
967m
1006m
1572m1567m
1956m
1500m
Montalto
Figure 4. Percentage of population killed by 1783 earthquakes in cities and villages of southernmost Calabria (from numbers of inhabitants and
casualties listed by Vivenzio 1783). Solid lines delimit regions with most gravely struck population centres (highest fatality rates, see text for intensity
estimates). Fatalities in the area covered by the map, which does not include epicentral areas of the March 1 and 28 shocks (Fig. 3), resulted mostly
from shocks on February 5 and 6. Dotted lines represent the isoseismals IX, X and X–XI of the February 7 event according to Boschi et al. (1995).
1783 Calabria seismic sequence 505
# 2001 RAS, GJI 147, 499–516
been the ‘centres’ of the earthquakes (Fig. 3) at a time when
the physical nature of earthquakes was unclear (e.g. Hamilton
1783; Vivenzio 1783; Dolomieu 1784; Baratta 1901; see also
Postpischl 1985; Boschi et al. 1995). Both earthquakes had
pronounced effects on the regional morphology. The first
triggered numerous landslides of awesome proportions (referred
to as ‘sconvolgimento’) that dammed rivers to create 215 lakes,
five of which were 1 km or more in length (Vivenzio 1783;
Cotecchia et al. 1969). The second triggered a large, 2 km long
rockslide along the Tyrrhenian coastal cliff as well as a small
tsunami (Baratta 1901). The first shock also caused the wide-
spread occurrence of sand fountains. It is probable that, as
noted by Dolomieu and Baratta, the exceptional size and
pervasiveness of such effects was partly a result of the regional
geology and of a notably rainy winter. Nevertheless, effects of
this type are usually characteristic of intensities of at least X–XI
(corresponding to magnitudes i6.5 using the relationship
between magnitude and epicentral intensity I0 for the Italian
peninsula M=0.53I0+0.96; Armijo et al. 1986). Hence, it is
probably reasonable to infer that the continuous contours
drawn in Fig. 4 roughly follow the isoseismals X and XI or XI
and XII. The surface area of the zone within the outer contour
(intensity X or XI isoseismal) of the largest lobe suggests,
in turn, that the magnitude of the first, main February 5
shock was at least equal to 7, as implied by its large environ-
mental effects, corresponding to intensities of XI–XII (see also
Postpischl 1985; Boschi et al. 1995). Similarly, the second,
February 6 shock probably reached an intensity of about
X–XI, corresponding to a magnitude of at least 6.5.
Like the first two events, the two subsequent events of the
1783 Calabrian earthquake sequence were probably shallow
crustal events. They were smaller than the main February 5
shock. The February 7 earthquake was clearly larger than
the March 1 shock. As implied by the contours of the meso-
seismic area (Fig. 3), the February 7 event was located in the
southeastern part of the Mesima Basin, at the western foot of
the Serre Mountains, where it reduced to heaps of ruins all the
villages between Acquaro and Soriano Calabro, and caused
destruction between Pizzoni and Dinami (a distance of 15 km;
Fig. 4), with landslides in the vicinity of Soriano, Arena and
Mileto (e.g. Dolomieu 1784; see also Boschi et al. 1995). Note
that since most of the inhabitants were camping out of their
houses at this time, the percentage of casualties cannot be used
to assess the size of the earthquake. The March 1 event was
located along the northernmost edge of the Mesima Basin,
at the northwestern foot of the Serre Mountains, where it
caused more limited damage (Fig. 3, see also Postpischl 1985;
Boschi et al. 1995). Thus, while the February 7 shock could
have reached an intensity of X or slightly more (Fig. 4, see
also Boschi et al. 1995) and a magnitude of at most 6.5,
that on March 1 probably had an intensity of IX at most
(e.g. Postpischl 1985) and a magnitude smaller than 6. For
the March 28 event, the mesoseismal area (Fig. 3) suggests a
location in the Catanzaro Basin, near Borgia, about 20 km
WNW of the March 1 event. It may have reached an intensity
of XI (according to Boschi et al. 1995) and a magnitude of the
order of, or somewhat greater than, 6.5. However, this shock
appears to have been deeper than the previous ones. The
gradient defined by the isoseismals, which is less steep than that
associated with the main event of February 5 (Fig. 3), and the
fact it was felt at greater distances (Baratta 1901) attest to a
deeper rupture nucleation.
RECENT AND ACT IVE FAULTS OFSOUTHERN CALABRIA
The main normal fault system separating the South Calabrian
Upper Pliocene–Pleistocene basins from the uplifted mountain
ranges comprises chiefly, from north to south, the Serre Fault,
along the west side of the Serre Mountains, the Galatro,
Cittanova (Santa Cristina) and Sant’Eufemia faults, along
the northwest edge of the Aspromonte massif, and the Scilla
Fault, bounding the Palmi high along the Tyrrhenian Sea
(Figs 1 and 2).
Given the Plio-Quaternary geological framework, the infer-
ence drawn from the macroseismic observations that the 1783
February 5 and 6 earthquakes had their hypocentres at shallow
depths just west of the Aspromonte and Palmi highs (Figs 1
and 3) implies that these two earthquakes resulted from slip
on the west-dipping normal faults bounding these two base-
ment highs. Similarly, the February 7 andMarch 1 earthquakes
resulted from slip on the west-dipping normal fault separating
the Serre basement high from the Mesima Basin.
Morphological and structural evidence confirms these con-
clusions and yields first-order constraints on the kinematics and
rates of active faulting in this southernmost part of Calabria.
Surface break of the 1783 February 5 earthquake on theCittanova Fault
Alone among the scholars and engineers who visited southern
Calabria after the 1783 earthquakes, Deodat Gratet de Dolomieu
was a dedicated Earth Scientist. His description of the meso-
seismal area of the Calabrian earthquakes (Dolomieu 1784)
places the morphological effects of these earthquakes in a
geological framework that is clear and correct by modern
standards. Although Dolomieu did not understand the origin
of the earthquakes at the time, his accurate observations throw
definitive light on surface deformation. Two paragraphs of his
‘Memoire’ (Dolomieu 1784, p. 36 and p. 46), which we quote
fully below, provide the key:
‘Il s’ensuivit, que dans presque toute la longueur de la chaine,
les terreins, qui etoient appuyes contre le granit de la base des
monts Caulone, Esope, Sagra et Aspramonte, glisserent sur ce
noyeau solide, dont la pente est rapide, et descendirent un peu
plus bas. Il s’etablit alors une fente de plusieurs pieds de large,
sur une longueur de 9 a 10 milles, entre le solide et le terrein
sabloneux; et cette fente regne, presque sans discontinuite,
depuis Saint George, en suivant le contour des bases, jusque
derriere Sainte Cristine.’
‘Tout le sol de la plaine, qui entoure Casalnovo, s’est affaisse.
Cet abaissement est surtout fort aparent, au dessus du bourg,
au pied de la montagne. Tous les terreins inclines, apuyes
contre cette meme montagne, ont glisse plus bas; en laissant,
entre le terrein mouvant et le solide, des fentes de plusieurs
pieds de large, qui s’etendent a trois, ou quatre milles.’
[‘Along almost the entire length of the range, the rocks
that were standing against the granite at the base of Mts
Caulone, Esope, Sagra and Aspramonte slipped on that rigid
core, whose slope is steep, and descended a little lower. Thus
was established a fissure several feet wide, over a length of
9–10 miles between the solid rock and the sandy terrane, and
this fissure prevails, almost without discontinuity, from San
Giorgio, following the contour of the mountain base, as far as
behind Santa Christina.’
506 E. Jacques et al.
# 2001 RAS, GJI 147, 499–516
‘The floor of the whole plain around Casalnuovo fell down.
Such downthrow is mostly apparent above the village, at the
foot of the mountain. All the sloping terranes that stood
against this mountain slid down, leaving, between the solid
rock and the moving terrane, open fissures several feet wide
over a length of 3–4 miles.’]
The length, continuity and position of the zone of down-
throw and fissures, along the interface between bedrock and
Quaternary sediments, leave little doubt that Dolomieu, while
interpreting this zone as the trace of an exceptionally large
landslide, linked to regional compaction, was in fact the first to
describe the surface break of an earthquake.
From Dolomieu’s observations, it was clearly the Cittanova
Fault that ruptured over a length of 18 km during the 1783
February 5 main shock. Seismic slip on the fault appears
to have been the largest between S. Giorgio and S. Cristina,
possibly with a maximum near Molochio. The presence of
open fissures at the base of the coseismic scarp is in keeping
with extensional faulting, with slip on the fault plane at depth
propagating more steeply near the surface, giving rise to cracks,
as commonly observed, for instance, in Afar (Jacques 1995).
The reported crack opening (several feet) was probably of the
order of 3–4 ft (>2), corresponding to a width=1 m (the foot
size used by Dolomieu was equal toy0.32 m). On a fault plane
dipping y70u near the surface, such a width would be con-
sistent with about 3 m slip. This motion clearly occurred along
the forested slope break still visible above the olive groves at
the range-front base near Molochio (Fig. 5a).
Geomorphological and structural evidence for activefaulting along the northwestern edge of the Aspromonteand Serre mountains
There is much structural and geomorphological evidence con-
firming recent, ongoing movement on the normal faults of
southern Calabria. The faults clearly cut and deform the Upper
Pleistocene beach sands deposits (Tortorici et al. 1995) and
control the regional topography (Figs 5 and 6) giving rise to
prominent range-front escarpments (Tapponnier et al. 1987),
along which other large earthquakes have destroyed towns and
villages in the past (e.g. 1659 and 1791, Fig. 1).
The 40 km long cumulative escarpment of the NNE–SSW-
trending, WNW-dipping Serre Fault is up to 200 m high
(Fig. 7a). The footwall basement rocks are overlain by two
terrace levels of Early Middle Pleistocene age (Ghisetti 1981).
Narrow V-shaped valleys separate well-developed triangular or
trapezoidal facets along the range front. The NNE-trending
Galatro Fault, within the left step between the Serre and
Cittanova faults, is only 6 km long, with a 150 m high cumu-
lative escarpment (Figs 1 and 7b). This short fault is probably
the northernmost segment of the y40 km long, NE-trending
Cittanova Fault, with which it may merge at depth.
Along the Cittanova Fault, the Upper Pleistocene sands
are strongly tilted westwards and cut by numerous secondary
faults, some of them rotated into reverse attitude (Fig. 5b). A
1–2 m wide gouge zone separates the sands from the crystalline
basement rocks. Mesostructural analysis of the gouge indicates
dip-slipmovement andN140uEt10u extension (Tapponnier et al.
1987; Tortorici et al. 1995). The morphology of the Cittanova
Fault escarpment varies along strike. The 10 km long south-
western stretch of the fault is marked by a cumulative scarp of
rather small height (100–150 m), while towards the northeast
this scarp (y30 km long) reaches a height of 400–700 m
(Figs 5a, 6 and 7c). This highest part of the escarpment is
shaped by three sets of triangular facets (Tortorici et al. 1995).
It is at the base of the lowest facets that Dolomieu described the
most prominent break of the main shock of the 1783 sequence
(Tapponnier et al. 1987), suggesting a relationship between
largest slip per event, fastest slip rate and greatest finite offset.
While the footwall to the southwest is overlain by only one
Lower Pleistocene terrace (Fig. 2), a second terrace level of
Mid-Pleistocene age (Ghisetti 1981; Dumas et al. 1982) exists in
the northeast (Figs 2a, 5a and 7c). Because Pleistocene marine
sedimentation reached the range front in the northeast, it is
likely that the different sets of facets and terrace levels there
correlate with climatic sea-level highstands. Stratigraphic and
geomorphological data suggest a minimum vertical slip rate
along the Cittanova Fault of 0.7 mm yrx1 in the last 125 kyr
(Tortorici et al. 1995). We interpret the smaller vertical throw
on the southwestern stretch of the Cittanova Fault as resulting
from a combination of strike change and slip transfer. The
more westerly strike of this stretch implies a sinistral com-
ponent of slip, as observed in the field (Tortorici et al. 1995)
(Fig. 1, inset). More importantly, extension is probably distri-
buted between this stretch and the nearby Sant’Eufemia Fault,
as commonly observed along overlapping normal fault systems
(e.g. Manighetti et al. 2001a,b).
West of the Delianuova right step, the 18 km long, ENE–
WSW-trending Sant’Eufemia Fault dips to the NNW, bounding
the southern sector of the Gioia Tauro Basin (Figs 1, 2 and 6).
The cumulative escarpment of this fault is up to 300 m high
(Fig. 7d). Its southwestern part, y10 km long, displays two sets
of triangular facets separated by wine-glass valleys (see Fig. 7
in Tortorici et al. 1995), with two distinct terrace levels of
Lower–Middle Pleistocene age on the uplifted footwall (Ghisetti
1981; Dumas et al. 1982). Shear planes in crystalline footwall
rocks show oblique slickensides (pitches of 35u–70u) consistentwith a left-lateral component of slip and a N135uEt5u exten-sion direction (Fig. 1, inset) (Tortorici et al. 1995). An uplift
rate of y0.7 mm yrx1 in the last 125 kyr, comparable to that
on the Cittanova Fault, has been estimated on this fault
(Tortorici et al. 1995).
Finally, the NW-dipping, NNE–SSW-trending Scilla Fault
bounds the Tyrrhenian shore of the Palmi high (Figs 1 and 2).
Southwest of Bagnara it veers to an ENE–WSW direction,
which it keeps onland to Villa San Giovanni. Between Bagnara
and Scilla, the escarpment forms a sea cliff up to 600 m high
(Fig. 7e). The crystalline rocks of the footwall are capped by a
Middle Pleistocene marine terrace, tilted towards the southeast
(Ghisetti 1981; Dumas et al. 1982). Southwest of Bagnara, up to
four levels of uplifted marine terraces (Figs 2 and 7e), reflecting
Pantosti 1992; Westaway 1993; Tortorici et al. 1995), appear to
correspond to isotopic stages 11, 9, 7 and 5 (Bassinot et al.
1994). East of Villa San Giovanni, the 90 m offset of the lower
level (stage 5) is again consistent with a throw rate of about
0.7 mm yrx1 in the last 125 kyr.
MODELL ING OF STAT IC STRESSCHANGES INDUCED BY THE 1 7 8 3SE I SMIC SEQUENCE
Following the overall approach taken by Reasenberg &
Simpson (1992) and King et al. (1994), we model the faults as
1783 Calabria seismic sequence 507
# 2001 RAS, GJI 147, 499–516
W E
CFTilte
d late Pleistocene beach sand Aspromonte
Basement
Gougezone
2 m
(5b)
(5a)
Figure 5. (a) Aspromonte front near Molochio (shown in Fig. 2 by a dotted box). Slope break at base of faceted spurs (arrows) corresponds to trace
of Cittanova (Santa Cristina) Fault and to part of the 1783 surface scarp mentioned by Dolomieu (1784). (b) Cittanova Fault zone (site indicated by an
arrow in Fig. 2). Normal masterfault (CF) separates footwall Aspromonte gneisses to the southeast (right-hand side of photomontage) from upper
Pleistocene Gioia Basin beach deposits in hangingwall (left-hand side). View to NE. Pleistocene deposits are strongly tilted (#N40uE, 35–40uNW) and
faulted. Several faults striking parallel to CF (N40uE) have steep eastward dips and appear to be normal faults tilted into reverse position.
508 E. Jacques et al.
# 2001 RAS, GJI 147, 499–516
discontinuities in an elastic half-space. To study the effect of
stress changes caused by the successive ruptures of different
fault segments during the earthquake sequence, we calculate
the changes of the Coulomb failure stress sf on either opti-
mally oriented fault planes or fault planes with given strikes
(Dsf=DtxmkDsn, where Dsn and Dt are changes in normal
and shear stresses on the fault planes and mk is the effective
friction coefficient). Following King et al. (1994), optimum
fault angles are determined using the stress field obtained by
adding a regional stress field to that created by the seismic
dislocations and calculating the directions that maximize sf onthose planes. The areas where sf increases are those where
subsequent brittle failure is promoted. Where sf drops, faultingis less probable.
Parameters used for Coulomb modelling
From focal depth determinations of the 1908 and 1975Messina
Strait earthquakes (Gasparini et al. 1982; Cello et al. 1982;
Bottari et al. 1989), we infer the regional seismogenic thickness
to be y20 km. This is compatible with the maximum magni-
tude (y7) of known crustal events. We thus take the maximum
depth of seismic faulting to be 20 km. We show the results of
Coulomb stress calculations at a depth of 10 km (mid-depth
of the seismogenic crust), using a standard shear modulus of
3.3r1010 Pa and a friction coefficient of 0.75. Given the size
of the first three events, which implies a y15 km depth of
nucleation, the 10 km depth choice seems adequate. For the
probably shallower 1783 March 1 My5.7 event (y6 km), a
depth of 5 km was chosen to assess the triggering effects of
previous shocks. In any case, Dsf is not very sensitive to depth,
except close to faults, where small uncertainties in parameters
such as dip, strike, slip distribution, fault length and sinuosity
can lead to significant changes. Similarly, the choice of the
modulus only changes the absolute values of sf. Errors in
estimating the seismic moment only change sf amplitudes. The
value of the friction coefficient, which is also not very critical
(King et al. 1994), is selected to be consistent with observations,
as discussed below.
In southwestern Calabria, the recent extension direction
inferred from slickenside measurements is about N125uE (Cello
et al. 1982; Tortorici et al. 1995), compatible with the fault
plane solutions of five out of seven shallow events (December
1908, January 1975, April 1978 and February 1980) (Cello et al.
1982; Gasparini et al. 1982; Anderson & Jackson 1987). Three
solutions show dominant dip slip on N10uE–N50uE-strikingnormal faults. The other two are dominantly strike-slip on
either N30uE–N60uE-striking left-lateral planes or N120uE–N150uE-striking right-lateral planes. Leaving aside the 1947
and 1990 events, located on the outer fringe of the region of
interest, we thus take a N125uE regional extension direction for
modelling. In general, regional stress magnitudes are poorly
known. We use a low value (40 bar) for the regional deviatoric
extensional stress s3. The results would not be significantly
altered by choosing a higher value. Moreover, as pointed out
for strike-slip faulting by King et al. (1994), the Dsf distributionis relatively insensitive to the amplitude of the regional stress.
Except very close to the faults, the friction coefficient and
stress direction chosen yield optimum left-lateral faults orientated
N70uE and normal faults striking N35uE. This is compatible
with the observed direction of left-lateral (N45uE–N90uE) andnormal faults (N25uE–N45uE) in the field (Tortorici et al.
1995). The relatively high value of the friction coefficient (0.75)
chosen yields an optimum dip of y63u for the optimal normal
fault, consistent with the dips observed in the cross-sections
[Figs 1 (inset), 2 and 6] (Tortorici et al. 1995).
Seismic slip during the 1783 February 5 event
The location of the main 1783 February 5 shock (Fig. 3)
and the corresponding fatality distribution (Fig. 4) imply that
it resulted from rupture of three fault segments (Cittanova,
Sant’Eufemia and Galatro). Given the length of surface rupture
and maximum slip observed by Dolomieu, it is likely that the
entire northeasternmost 30 km of the Cittanova Fault, near
which the most extensive destruction occurred, slipped during
the main shock. The Sant’Eufemia Fault, along which damage
and casualties were similarly great on February 5, probably
slipped at about the same time over a minimum length of
y10 km. The 6 km long Galatro Fault, which extends beyond
the NE end of the Cittanova Fault did not generate a separate
event, but probably also slipped at the time of the main shock.
This interpretation brings the total length of faulting during the
February 5 event to y45 km. With a mean fault dip of y65u,the overall rupture area is about 1000 km2. The empirical magni-
tude versus rupture area relationship of Wells & Coppersmith
Cittanova Fault S. Eufemia Fault
Figure 6. General view of uplifted basement steps along the Sant’Eufemia and Cittanova (Santa Cristina) faults (looking SSE from Sant’Elia
Mountain near Palmi), triangular dotted line box in Fig. 2. Gioia Tauro Basin is in the foreground.
1783 Calabria seismic sequence 509
# 2001 RAS, GJI 147, 499–516
(1994) would thus imply a magnitude ofy7 for this shock. Other
scaling laws (Scholz 1990; Jacques 1995) would yield a magni-
tude of about 7.2, or a seismic moment, M0, of y7r1019 Nm.
All these estimates are in agreement with the magnitude inferred
from discussion of the mesoseismal macroseismic effects. Finally,
an Mw magnitude ofy7.2 would imply an average slip ofy2 m
using the magnitude relationship Mw=(2 /3)log(mSDu)x6.03
of Kanamori (1977). This is also consistent with the maximum
slip of y3 m observed on the Cittanova Fault by Dolomieu
(1784), and with the empirical relationship between maximum
slip and magnitude of Wells & Coppersmith (1994).
Instead of using the same average slip value on the three fault
segments ruptured by the February 5 main shock, one may use
long-term geomorphological evidence to estimate the relative
amplitudes of average seismic slip on the Galatro, Sant’Eufemia
and Cittanova faults. We assume the average seismic slip (Du)
to be linearly related to the apparent cumulative vertical offset
(Cvo) on each fault (King et al. 1988) (Du=aCvo+b). a and b
can be derived from the values observed along the Cittanova
Fault (Duy2.5 m, between the average of 2 m and the maxi-
mum value of 3 m; Cvo=440 m) and from the condition that
the total moment resulting from coseismic slip on the three
faults is y7r1019 N m (Table 2). We obtain a=0.0054 and
b=0.0436 m. We also assume a uniform orientation of the slip
vector on the three faults, which yields dip values of 60u, 65uand 70u for the Cittanova, Galatro and Sant’Eufemia faults,
SW
0200400600800
1000
0 5
altit
ude
(m)
Projection distance along fault trace (km)
NE
(e)Scilla fault
(d)S. Eufemia faultWSW ENE
(c)Cittanova faultSW
(b)Galatro faultSSW NNE
(a)Serre faultSSW NNE
0 5 10 15 20 25
0200400600800
altit
ude
(m)
Projection distance along fault trace (km)
0200400600800
10001200
0 5 10 15 20 25
altit
ude
(m)
Projection distance along fault trace (km)
0200400600800
100012001400
0 5 10 15 20 25 30 35 40 45
altit
ude
(m)
Projection distance along fault trace (km)
0200400600800
1000
0 5 10 15 20 25 30 35 40
altit
ude
(m)
Projection distance along fault trace (km)
Figure 7. Projections, on appropriately oriented vertical planes, of topographic profiles along (a) the Serre (N30uE), (b) the Galatro (N20uE),(c) the Cittanova (N40uE), (d) the Sant’Eufemia (N70uE) and (e) the Scilla (N45uE) fault traces (crosses), and along the outer edges of terrace levels
(open squares, grey dots and black diamonds) preserved on footwalls (from IGM 1:25 000 topographic maps). Vertical exaggeration is 4. Minimum
cumulative vertical offset is between the upper terrace level (black diamonds) and the fault trace (crosses).
510 E. Jacques et al.
# 2001 RAS, GJI 147, 499–516
respectively (Table 2). Given the mean strike of the Cittanova
Fault (N40uE), which is nearly orthogonal to the regional
extension direction (N125uE), consistent with pure normal fault-
ing, this kinematic compatibility condition also yields pitches
of 95uN and 75uW on the Galatro and Sant’Eufemia faults,
respectively, in agreement with field observations (Fig. 1, inset)
(Tortorici et al. 1995). The final parameters used as inputs for
the Coulomb modelling on each fault segment are given in
Table 2.
Coulomb stress changes caused by the 1783 February 5main shock on the fault of the 1783 February 6 event
The main shock increased sf by more than 10 bar both north of
Galatro and west of Sant’Eufemia, whether on normal faults
striking N35uE and N70uE (Fig. 8a) or on optimally oriented
strike-slip faults (left-lateral, striking yN70uE, Fig. 8b).Thus, the February 5 earthquake most probably promoted
rupturing of the N70uE-trending, southwest part of the Scilla
Fault. Considering only static stress changes, the rupture might
have stopped near Bagnara (where the fault bends to a NNE
direction), but macroseismic observations (Fig. 4) suggest that
it propagated northeastwards to Palmi.
Seismic slip during the 1783 February 6 event
According to models of circular rupture (e.g. Scholz 1990) and
assuming a stress drop of 30 bar (an average value, Kanamori
1977), the rupture diameters of M=6–6.5 earthquakes range
between 11 and 19 km. With a mean dip of about 65u, suchruptures would cut the crust to depths of 10–17 km, falling
short of the base of the seismogenic layer, in keeping with the
use of such models.
The length of the Scilla Fault lying within the area of high
sf increase is about 14 km, suggesting a minimum equivalent
circular rupture surface of 16 km diameter. This corresponds
to a minimum seismic moment of y3.7r1018 Nm (average
minimum Du=0.55 m) and a minimum magnitude Mw ofy6.3.
This lower bound is compatible with the magnitude (i6.5)
derived from macroseismic observations, which suggest that the
rupture propagated to Palmi over a total length of y20 km. To
examine the effects of static stress change, we use a square
rupture with a side length of 14 km, keeping in mind that such
effects are minimized by this choice.
Because the macroseismic effects of the February 6 shock
include a tidal wave, we assume that the rupture reached the
surface. We take the fault dip to be y65u and the pitch or rake
of the slip vector to be identical to that on the Sant’Eufemia
Fault, almost parallel to the trace of the southwest part of the
Scilla Fault.
Coulomb stress changes caused by the February 5 and 6events on the fault of the 1783 February 7 event and inthe Messina Strait
After the February 5 and 6 shocks, sf rose by more than
1–5 bar on the southwest part of the Serre Fault where the
February 7 event took place (Fig. 8c). This fault is nearly
parallel to optimally oriented normal faults, suggesting that
the combined effects of the first two events promoted the
occurrence of the February 7 shock.
Another, smaller event occurred in the evening of February 7
(22:00 local time), about 2 hr after the third principal shock of
the sequence, and was destructive mostly at Messina and to a
lesser degree at Sant’Agata di Reggio. This earthquake was
probably located in the Messina Strait region (Baratta 1901)
within the area where the Coulomb stress rose by 1–5 bar south
of the Scilla and Sant’Eufemia faults (Figs 1, 8c and 8d).
Seismic slip during the 1783 February 7 and 1783March 1 events
The macroseismic effects of the February 7 event are con-
sistent with a magnitude of y6.5 and a seismic moment of
y6.2r1018 Nm. Using the circular rupture model and the
same assumption on stress drop as for the February 6 event, we
approximate the source by a square-shaped rupture with a side
length of 17 km centred on the area of maximum destruction.
As for the February 6 shock we assume a fault dip of 65u, andthat the rupture nearly reached the surface. The segment of
the Serre Fault activated by the February 7 earthquake (average
Du=0.65 m) lies fully within the area where sf increased
between 1 and 10 bar due to the previous 1783 events (Fig. 8c).
However, between the Galatro Fault, which we infer to have
slipped during the main shock, and the 17 km long segment of
the Serre Fault broken by the February 7 event, there still exists
a 4 km long stretch of the Serre Fault that seems not to have
ruptured (Figs 1 and 8d) although it lies in an area where sfrose by 5–10 bar. Other than an unreported fore- or aftershock,
or aseismic strain relaxation, one possible explanation for
this apparent rupture gap is that we have underestimated the
rupture length of the February 7 event.
The 1783 March 1 event appears to have also ruptured
the Serre Fault further north, but its macroseismic effects
imply a magnitude of only y5.7 and a seismic moment of
y3.9r1017 Nm. Using the same assumptions as for the
February 6 and 7 events, its source parameters (including
an average Du=0.30 m) may be approximated by a shallow
square-shaped rupture surface with a side length of y7 km
centred on the area of maximum destruction with a dip of 65uand reaching a depth of y6 km below the surface.
Coulomb stress changes caused by previous 1783 eventson the fault of the March 1 event
Fig. 8(d) shows the Dsf caused by the first three shocks of the
1783 sequence computed in this case on optimally oriented
normal faults at a depth of 5 km, near the bottom of the
inferred source of the March 1 event, consistent with the usual
nucleation depth of M>5 earthquakes. That the fault segment
that activated the March 1 earthquake is oriented nearly
parallel to the optimal normal faults and lies in an area where sf
Table 2. Fault parameters used in Coulomb modelling.
Figures 8. Coulomb stress changes (Dsf) calculated after the main February 5 shock at a depth of 10 km (a) on N70uE-striking normal faults and
(b) on optimally oriented strike-slip faults (regional s3yN125uE). Ruptured fault segments are drawn in white. The trace, at 10 km depth, of the Scilla
Fault is drawn as a black dashed line. From (a) it appears that failure on the Scilla Fault is promoted, and from (b) a component of sinistral strike-slip
is expected. (c) Coulomb stress changes (Dsf) calculated after the February 5 and 6 earthquakes on optimally oriented normal faults (ruptured fault
segments are drawn in white). The trace of the southern part of the Serre Fault, at 10 km depth, is drawn as a black dashed line. It appears that failure
on the southern part of the Serre Fault (1783 February 7 event) is promoted (Coulomb stress increase from 1 to 5 bar). (d) Coulomb stress changes
(Dsf) calculated at 5 km depth after the February 5, 6 and 7 earthquakes on optimally oriented normal faults (ruptured faults segments drawn in
white). The trace of the northern part of the Serre Fault, at 5 km depth, is drawn as a black dashed line. It appears that the Coulomb stress increased
by about 1 bar on the northern part of the Serre Fault (1783 March 1 shock). The 1659 November 5 event is plotted on this figure as a black-circled
blue dot. Coulomb stress changes (Dsf) calculated after the February 5, 6, 7 and March 1 earthquakes (e) on optimally oriented normal faults and
(f) on optimally oriented strike-slip faults at 10 km depth. The 1783 March 28 event is plotted on these figures as a black-circled white dot. Note that
the 1908 December 28 earthquake (white-circled black dot) located in the Messina Strait lies in a area where sf increased by about 2 bar on SSE-
striking normal faults. The 1659 November 5 event is plotted on these figures as a black-circled blue dot; it could have previously ruptured the central
segment of the Serre Fault.
# 2001 RAS, GJI 147, 499–516
rose byy1 bar after the February 7 event is consistent with the
idea that this event was triggered by the first three events of
the 1783 sequence.
Effects of Coulomb stress changes caused by the firstfour shocks of the 1783 Calabria sequence and by the1659 and 1693 events on subsequent earthquakes
Although within a zone of strong sf increase (1–5 bar)
following the February 7 event, the central portion of the
Serre Fault apparently remained a rupture gap in 1783. We
suggest that this is because that stretch of the fault had already
ruptured 124 yr previously. Reports of damage and casualties,
which were extensive and serious (Baratta 1901), show that
the 1659 November 5 earthquake (I0=X, Boschi et al. 1995;
corresponding to My6) (Figs 1, 8e and 8f) was probably due
to slip on the central Serre Fault. This event probably ruptured
the middle 10 km long segment of the fault, causing on it a
decrease of sf that remained large enough, over a century later,
to prevent rupture.
Figs 8(e) and (f) show the Dsf caused by the first four
earthquakes of the 1783 sequence, at 10 km depth, on optimally
oriented normal or strike-slip faults. The largest increases, with
Dsf reaching more than 10 bar, occur south of the Scilla and
Sant’Eufemia faults and in the central overstep between the
Cittanova, Galatro and Serre faults. Along the northern part of
the Serre Fault, the sf increase is only slightly over 5 bar.
Including the My6 1659 earthquake in the Coulomb
modelling, with assumptions similar to those used for the
1783 February 7 shock, augments the sf increase in the region
immediately west of the epicentre of the 1783 March 28 shock,
the last of the sequence, near Borgia, south of the Catanzaro
Basin (Figs 1 and 9). Assuming that this event nucleated in
the upper crust on an optimally oriented normal or strike-slip
fault, the sf increase would be only y0.3 bar without the 1659
event (Figs 8e and f) and between 0.6 and 1 bar with it (Fig. 9).
This could account for the triggering of the last event, possibly
along a left-lateral strike-slip fault orientated WSW–ENE, in
keeping with the maximum elongation of the mesoseismal area
(Fig. 3).
The trace and nature of the fault that slipped during this last
event remain unclear, however, and the hypocentre depth
uncertain, although probably deeper than that of the others. If,
as has been inferred from macroseismic evidence reported by
Boschi et al. (1995), the 1783 March 28 shock had a magnitude
of y6.8 and a seismic moment of y1.8r1019 N m, the circular
rupture model would yield a dislocation area 28 km across,
implying that this event, if shallow, should have cut through
the seismogenic crust, causing clear surface breaks. The lack
of evidence for coseismic surface faulting is thus in keeping
with deeper nucleation. All this prevents us from drawing
firmer conclusions about the triggering of this event, and from
introducing it in further Coulomb modelling.
One last, minor shock of the sequence, followed by an
aftershock swarm, occurred on April 26 1783 (Baratta 1901)